Listeria monocytogenes 8

125

From: Bacterial Genomes and Infectious Diseases

Edited by: V. L. Chan, P. M. Sherman, and B. Bourke © Humana Press Inc., Totowa, NJ

8

Listeria monocytogenes

Keith Ireton

Summary

Listeria monocytogenes is a Gram-positive, intracellular bacterial pathogen responsible for severe food-borne illnesses resulting in central nervous system infection or abortion. L. monocytogenes induces its own internalization into mammalian cells, escapes from the host cell phagosome, repli- cates extensively in the cytosol, and spreads from one host cell to another through an F-actin-depen- dent motility process. Previously, classical genetic approaches were used to identify bacterial virulence factors critical for the intracellular life cycle of L. monocytogenes. The recent availabil- ity of the nucleotide sequence of L. monocytogenes has provided the potential for global analysis of bacterial proteins that affect pathogenesis. In this chapter, ways in which the L. monocytogenes genome has been used to probe the functions of bacterial proteins in virulence is discussed. At the end of the chapter, future genomic- or proteomic-based approaches that might improve or expand on current work are highlighted.

Key Words: Genomics; internalins; leucine-rich repeat; Listeria monocytogenes; listeriosis; PrfA;

lipoprotein; LPXTG; sortase; surface protein.

1. Introduction

Listeria monocytogenes is a Gram-positive, food-borne pathogen capable of causing gastroenteritis and severe systemic infections culminating in meningitis or abortion (1). The incidence of listeriosis in the general population is low (~2500 cases per year in the United States), representing less than 0.1% of all food-borne illnesses (2). How- ever, the high mortality rate of listeriosis (~20%) makes L. monocytogenes responsible for almost 30% of deaths caused by food-borne pathogens.

L. monocytogenes is a facultative, intracellular pathogen that occupies several environ- mental niches (1). This bacterium infects a wide variety of animal hosts in addition to humans, including other mammals (sheep, cattle, goats, pigs, rabbits, mice), birds, and fish. Moreover, L. monocytogenes replicates outside of host animals, and is thought to live in the soil as a saprophyte utilizing decaying vegetation as a food source. Thus, in contrast to obligate bacterial pathogens of humans, such as Neisseria spp. or Chlamy- dia spp., L. monocytogenes exercises multiple lifestyles that are adapted for different environmental situations. Consistent with this greater breadth of habitat, the genome size of L. monocytogenes (~3 Mb [3]) is larger than those of the aforementioned obli- gate pathogens (4–6).

Much of our current knowledge of the functions of individual genes in L. monocytogenes

has focused on those involved in virulence in animal models, and probably humans

(1,7). Originally, classic “phenotype-driven” genetic screens were performed to identify bacterial virulence genes (1), and genetic, biochemical, or immunological approaches were employed to define mouse or human proteins that influence host susceptibility to infection (8). The recent availability of the nucleotide sequences of the genomes of L.

monocytogenes and its human or animal hosts provides the potential for global analy- sis of bacterial and host factors that influence pathogenesis. This chapter begins with a brief introduction to the intracellular life cycle of L. monocytogenes, and the functions of virulence factors identified through classic genetic screens (Subheading 8.2.). Then, the general characteristics of the L. monocytogenes genome are described (Subhead- ing 8.3.), followed by a discussion on how DNA sequence information has prompted genetic analysis of the roles of particular bacterial genes in virulence (Subheading 8.4.).

Recent genomic-based studies of bacterial gene expression responses are covered in Subheading 8.5. Finally, at the end of the chapter, future genomic- or proteomic- based approaches that might improve or expand on current work are highlighted (Sub- heading 8.6.).

2. Intracellular Life Cycle of L. monocytogenes

L. monocytogenes has the ability to penetrate host mammalian cells, to replicate within cells, and to spread from one cell to another while always remaining in the host cyto- sol. The various stages of the L. monocytogenes intracellular life cycle were originally characterized through transmission electron microscopy analysis of infected cultured mammalian cells (9). Genetic approaches led to the identification of bacterial factors controlling the different steps in the life cycle (1). A cartoon based on the seminal elec- tron microscopy work by Tilney and Portnoy is presented in Fig. 1, with virulence pro- teins indicated next to the step(s) that they regulate. The life cycle is briefly described in Subheading 2.1. For a detailed discussion of molecular mechanisms of Listeria virul- ence proteins, the reader is referred to an excellent comprehensive review by Vazquez- Boland and coauthors (1).

Fig. 1. The intracellular life cycle of Listeria monocytogenes. Bacterial virulence factors are

indicated next to the stage of infection that they control. (Adapted from ref. 9, with permission

from The Rockefeller University Press.)

2.1. Bacterial Entry, Phagosome Dissolution, and Cell–Cell Spread

On interaction with host mammalian cells, L. monocytogenes is rapidly internalized into membranous structures called phagosomes. Listeria is engulfed by professional phagocytes, such as macrophages or neutrophils, and also by cells that are generally thought of as nonphagocytic, including epithelial and endothelial cells (1). Internaliza- tion (“entry”) into nonphagocytic cells requires active participation of both the bacte- rium and the host cell. The two major bacterial factors that mediate Listeria entry into mammalian cells are the surface proteins Internalin A (InlA) and InlB (10). Each of these microbial proteins contains an amino-terminal leucine-rich repeat (LRR) domain that folds into a horseshoe-like structure, followed by an internal immunoglobulin-like inter- repeat (IR) region, and a carboxyl-terminal region that mediates cell surface anchoring (10,11). The LRR regions in InlA and InlB interact with distinct mammalian surface receptors, the cell–cell adhesion molecule, E-cadherin (12), or the receptor tyrosine kinase, Met (13), respectively. Bacterial engagement of E-cadherin or Met promote entry through modulation of mammalian signaling pathways that lead to F-actin-driven remodeling of the host cell surface (10,14,15).

Within approx 30 min after internalization, L. monocytogenes lyses the phagosome through the action of the secreted pore-forming hemolysin, listeriolysin O (LLO) (16).

In some mammalian cell lines, the secreted phospholipases, PlcA and PlcB, aid LLO in mediating phagosomal escape (17). Once free in the cytosol, pathogens replicate exten- sively and use the bacterial surface protein, ActA, to recruit host actin into filaments that become organized into tail-like structures (1,16). These “comet tails” propel patho- gens through the cytosol, and allow cell–cell spread through formation of a protrusion that is engulfed by an adjacent cell. The engulfed bacterium is found in a double-mem- branous, “secondary vacuole,” which it rapidly dissolves through the concerted action of LLO, PlcA, and PlcB (18,19). Having again accessed the cytosol, L. monocytogenes repeats its cycle of replication, motility, and cell–cell spread. With the exception of inlA and inlB, expression of the genes encoding each of the previously described viru- lence proteins requires the bacterial transcription factor PrfA (20,21). PrfA directly activates transcription of promoters of these genes by binding to an upstream sequence called a PrfA box. PrfA activity is upregulated in the phagosome and cytosol of host cells, ensuring that maximal synthesis of virulence proteins occurs in the intracellular environment in which they exert their functions.

2.2. Role of Bacterial Virulence Genes in Listeriosis

The intracellular life cycle of L. monocytogenes observed in cultured cells is thought to play an essential role in colonization of host tissues during human listeriosis (1,10, 22). Gastroenteritis likely arises from damage resulting from L. monocytogenes internal- ization and proliferation within cells lining the lumen of the small intestine, with a sub- sequent inflammatory response. In systemic infections, bacterial traversal of the intestinal cell barrier is probably followed by replication in macrophages and possibly hepato- cytes, with blood-borne spread to the central nervous systems (CNS) or placenta (1).

Work with animal models indicate that the bacterial genes depicted in Fig. 1 play key

roles in virulence. InlA/E–cadherin interaction appears to mediate translocation across

the intestinal cell barrier, based on experiments with guinea pigs or transgenic mice

expressing human E-cadherin (23). InlB is needed for infection of hepatocytes and

efficient colonization of the liver in mice (24,25). Bacterial propagation in the mouse liver, spleen, and other organs also requires LLO, ActA, and PrfA (1), indicating critical roles for intracellular replication and cell–cell spread. In contrast to these virulence fac- tors, PlcA and PlcB have redundant functions in colonization, which reflect the over- lapping roles for these phospholipases in mediating escape from primary and secondary phagosomes (18).

In addition to promoting infection of the liver, InlB might also contribute to CNS infections by inducing internalization into brain microvascular endothelial cells (1,22).

Alternatively, or additionally, infections of the CNS could be initiated by ActA-mediated spread of bacteria inside phagocytes to endothelium, or by transmigration of infected leukocytes across the endothelial barrier (22,26). Most abortions are thought to arise from L. monocytogenes infection of placental trophoblast cells, followed by entry into the fetal blood stream. Although InlA and InlB promote internalization into primary human tro- phoblast cultures or placental villous explants (27,28), neither of these bacterial surface proteins play a detectable role in traversal of the fetal–placental barrier in pregnant guinea pigs (28). These findings suggest that infection of trophoblasts in vivo may occur primar- ily through ActA-mediated cell–cell spread from maternal macrophages.

3. Listeria Genomes

3.1. L. monocytogenes Strain, EGDe

and the Nonpathogenic L. innocua Strain, CLIP11262

Although phenotype-driven genetic screens were instrumental in the identification and characterization of the L. monocytogenes virulence factors depicted in Fig. 1 (1,7), the determination of the complete sequence of the L. monocytogenes genome in 2001 ([3] http://genolist.pasteur.fr/ListiList/genome.cgi) has significantly impacted the way in which gene function is being studied. Many laboratories have used DNA sequence information to provide clues as to the potential role of particular genes in pathogenesis.

One trend has been to identify genes encoding proteins with sequence similarity to known virulence factors or proteins that are predicted to regulate the localization or activity of previously identified virulence factors. Another strategy, termed “comparative geno- mics” takes advantage of the availability of the genome sequence of the nonpathogenic Listeria species, L. innocua (3). The logic behind this strategy is that genes present in L. monocytogenes, but absent in L. innocua, might be involved in pathogenesis.

The genomes of the L. monocytogenes strain EGDe (serovar 1/2a) and the L. innocua strain CLIP11262 (serovar 6A) are similar in size (~3 Mb), overall guanine and cyto- sine content (~40%), gene identity, and gene organization. About 90% of the genes pres- ent in L. monocytogenes EGDe are also found in L. innocua, and approx 95% of the genes found in L. innocua are conserved in L. monocytogenes. These findings indicate that the difference in pathogenicity between these two Listeria species is owing to a rela- tively small number of genes representing a minor proportion of the genome.

A subset of the 270 EGDe-specific genes are likely to contribute to pathogenesis (3,7).

Among these are the known L. monocytogenes virulence genes depicted in Fig. 1. In

addition to InlA, InlB, and ActA, 26 of the EGDe-specific genes encode putative cell

surface proteins. Strikingly, 9 of the 26 surface-anchored proteins contain LRR and IR

domains that have sequence similarity to those in InlA and InlB (3,29,30). In addition

to these surface proteins, three putative secreted proteins with LRR and IR domains are

present in EGDe and absent from L. innocua (31,32). Together with InlA and InlB, these surface and secreted LRR proteins comprise the internalin family (11,33). LRR proteins are widespread in nature, being found in prokaryotic and eukaryotic proteins of diverse function (34). A feature common to virtually all LRR domains is that they mediate protein–protein interactions. However, without exception, different LRR pro- teins have distinct ligands, indicating that the “horseshoe” structure common to these proteins serves as a scaffold on which specific amino acid clusters are presented for interaction. The multitude of EGDe-specific LRR proteins suggests that internalins, in addition to InlA and InlB, likely play important roles in pathogenesis. This notion has already been confirmed for a small proportion of these L. monocytogenes LRR pro- teins (30,31,35). Apart from internalins and other cell surface or secreted factors, other classes of genes present in EGDe and absent in L. innocua include those mediating trans- port or metabolism of carbohydrates or other nutrients, and transcription factors.

3.2. Other L. monocytogenes Strains

The genome of EGDe was sequenced because it is arguably the most commonly used strain for molecular virulence and immunological studies. However, EGDe was isolated from an outbreak in rabbits (1) and it is not clear if this strain would be highly virulent for humans. L. monocytogenes strains are classified into 13 serotypes, which fall into distinct evolutionary lineages (36,37). Serotypes 1/2a, 1/2c, 3a, and 3c are in lineage I;

4b, 4d, 4e, 1/2b, and 3b serotypes occupy lineage II; and 4a and 4c strains comprise lineage III. Interestingly, certain serotypes appear to have greater pathogenic potential than others (38). The majority of epidemic and sporadic cases are associated with 4b strains, despite the fact that 1/2a and 1/2c strains are commonly found in food and the environment (36). Based on this data, the 1/2a strain EGDe probably does not have the full pathogenic potential manifested by 4b strains isolated from food outbreaks.

In an effort to identify bacterial proteins that might contribute to human epidemics, two different groups determined the nucleotide sequences of the genomes of three sero- type 4b strains associated with listeriosis outbreaks, and one 1/2a strain responsible for a sporadic illness (36,37). Comparison of genes present in the newly sequenced 1/2a strain and EGDe, and the three 4b strains revealed that individual strains of L. monocytogenes are surprisingly genetically diverse, with 2–5% (50–140) of genes being strain-spe- cific. This is approximately the same level of divergence observed between EGDe and L. innocua CLIP11262. In one of the studies, DNA microarrays containing EGDe genes were used to screen 93 different L. monocytogenes strains of various serotypes, includ- ing 25 1/2a and 22 4b strains (36). As expected, genes encoding the major virulence factors characterized in EGDe (InlA, InlB, LLO, ActA, PlcA, PlcB) were conserved among all of the 1/2a or 4b strains analyzed. These findings support the idea that these proteins constitute “core” virulence factors critical for pathogenesis in humans.

The study also led to a more precise determination of genes specific to 1/2a or 4b

strains (36). Nineteen genes present in all 1/2a strains and absent in all 4b strains were

found. Six of these genes encode potential surface proteins, four of which also contain

LRR domains likely to have structures similar to those of InlA and InlB. Transport pro-

teins represent another class of 1/2a-specific genes, with a putative oligopeptide perme-

ase ABC transporter and four phosphoenol pyruvate-dependent transport system enzyme

II components being absent in 4b strains. Eight genes found in 4b strains, but absent in

1/2a strains, were identified in the microarray study (36). Similar to the situation with 1/2a-specific factors, three of the 4b-specific proteins are surface-anchored LRR pro- teins. Other proteins present specifically in 4b strains include putative transcription factors and proteins whose function cannot be easily predicted. Based on the microarray data, it would appear that some of the differences in pathogenicity thought to exist between 1/2a and 4b strains is because of a distinct repertoire of LRR and other surface proteins.

3.3. Identification of Virulence Genes Through Comparative Genomics

Given that elucidation of the genomes of 4b strains is very recent, the roles of 4b- specific genes in pathogenesis have yet to be addressed. In contrast, use of the EGDe and L. innocua CLIP11262 genomes has already led to the identification of several bacte- rial genes that affect virulence in animal models. In Subheading 8.4., examples in which EGDe-specific proteins belonging to some of these categories were found to play impor- tant roles in pathogenesis are discussed. It should be noted, however, that the presence of a particular gene in both Listeria species does not exclude the possibility that it might encode a virulence factor, particularly if the corresponding protein participates in fun- damental processes, such as protein secretion or anchoring. Examples of genes con- served among Listeria species that contribute to L. monocytogenes pathogenesis are also presented.

4. Use of Genomics to Identify L. monocytogenes Virulence Proteins 4.1. Surface or Secreted Proteins

4.1.1. Translocation Across the Plasma Membrane

With the exception of the transcription factor PrfA, all of the well-studied Listeria virulence factors presented in Fig. 1 are anchored to the bacterial surface or secreted into the external environment, where they interact with host cell membranes or pro- teins. Each of these Listeria proteins contains an amino-terminal signal peptide (SP) sequence typical of substrates of the classical Sec-dependent secretion pathway (39).

The genome of L. monocytogenes EGDe encodes 219 SP-containing proteins (3), indi- cating that at least 8% of all proteins are likely to be exported through a Sec apparatus.

In Escherichia coli, transport through the Sec pathway is driven by the SecA ATPase, which provides energy for translocation and also presents SP-containing proteins to an integral membrane channel composed, in part, of SecY and SecE. E. coli and most other bacteria have single copies of these sec genes. Interestingly, L. monocytogenes and L.

innocua encode two SecA homologs, SecA1 and SecA2 (40). SecA1 is probably essen-

tial (40), and is thought to be needed for export of the majority of SP-containing proteins,

including InlA, InlB, LLO, PlcA, PlcB, and ActA. SecA2 is nonessential, required for

virulence in mice, and appears to mediate secretion of a smaller subset of proteins,

including the cell wall hydrolases p60 and N-acetylmuramidase, the fibronectin-bind-

ing protein/adhesin, FpbA, and several lipoproteins (41,42). Surprisingly, in contrast

to SecA1 substrates, some proteins that are exported in a SecA2-dependent manner

lack recognizable SP sequences (41,42). It is not clear how these secreted proteins are

recognized by the SecA2 apparatus.

4.1.2. Anchoring of Surface Proteins

After export by the Sec machinery, proteins are either localized to the cell surface or released into the extracellular milieu. Release (secretion) appears to be a default path- way that occurs when a particular protein lacks an anchoring region. There are four dif- ferent mechanisms, defined by the nature of the anchoring motif, that mediate surface association of proteins in Listeria species (33).

4.2. Hydrophobic Membrane Anchor

Some surface proteins have a carboxyl-terminal hydrophobic region that is inserted in the cytoplasmic membrane, followed by a positively charged, cytosolic “stop trans- fer” sequence. The protein’s amino-terminus projects through the peptidoglycan (33).

ActA is anchored to the bacterial surface through such a hydrophobic tail-based mecha- nism, and examination of the EGDe genome identified 10 additional proteins that are likely to have hydrophobic sequences embedded in the cytoplasmic membrane (33). The roles of these other surface proteins in virulence remain to be determined.

4.3. Linkage to the Cell Wall

A second general anchoring mechanism occurs through covalent linkage to the pep- tidoglycan. Many surface proteins of various Gram-positive bacterial species contain a carboxyl-terminal LPXTG sequence (where X is any amino acid) that undergoes enzy- matic cleavage between the threonine and glycine residues, followed by linkage of the carboxyl group of the threonine to meso-diaminopimelic acid in the cell wall (43). The first protein to be demonstrated to be anchored by such a mechanism was Staphyloccus aureus protein A. There are 17–21 additional LPXTG-containing proteins in S. aureus strains. The staphylococcal enzyme that promotes anchoring of LPXTG sequences is called Sortase A (43). S. aureus also encodes a related enzyme, Sortase B, which medi- ates cell wall linkage of the carboxyl-terminal motif, NPQTN. In contrast to Sortase A, staphylococcal Sortase B may have only one substrate, IsdC (43,44).

Genes encoding proteins similar to Sortases A and B were recently identified in the EGDe and L. innocua genomes (45–47). Like S. aureus, Listeria species exhibit a dis- proportionality in Sortase A and Sortase B substrates. L. monocytogenes InlA contains an LPXTG motif that mediates its cell surface localization (48). In addition to InlA, an astounding 40 other LPXTG-containing proteins are encoded in the EGDe genome (3,33). This number represents about 30% of all surface proteins in this strain. Nineteen of these LPXTG proteins contain LRR domains, and 11 of the LPXTG/LRR proteins are absent from L. innocua (3,29,30). L. monocytogenes Sortase B is thought to recog- nize a carboxyl-terminal NXXTN sequence, and might have only two substrates, SvpA and Lmo2186 (47,49). Both of these Listeria proteins exhibit significant amino acid similarity to staphylococcal IsdC (47).

In S. aureus, Sortase A plays an important role in virulence (50), underscoring the importance of LPXTG proteins in colonization of host tissues by this pathogen. In L.

monocytogenes, inactivation of srtA caused the expected elimination in cell surface

anchoring of InlA and several other LPXTG-containing proteins, and abolished InlA-

dependent entry into host cells (45). The srtA deletion also resulted in a virulence defect

in intravenously or orally inoculated mice. It is noteworthy that the srtA mutant was

more attenuated than an inlA deletion mutant, which does not exhibit a virulence pheno- type unless mice are genetically modified to express human E-cadherin (23,24). Taken together, these results indicate that, in addition to InlA, other LPXTG-containing pro- teins influence pathogenesis by acting at a step subsequent to E-cadherin-mediated tra- versal of the intestinal cell barrier.

In S. aureus, Sortase B is dispensable for initial host colonization, and is instead needed for persistence in host tissues (44). Sortase B and its substrate IsdC promote heme–iron acquisition, and it is thought that the virulence defect of srtB mutants may be because of impaired scavenging of iron in the host (43,51). Despite the fact that the putative L. monocytogenes substrates, Lmo2186 and SvpA, have sequence similarity to IsdC (47), these proteins do not have a detectable role in heme–iron utilization in L.

monocytogenes (49). Moreover, deletion of svpA or srtB does not substantially alter the median lethal dose (LD50) in intravenously inoculated mice (47,49), suggesting that surface anchoring by SrtB does not play an essential role in Listeria pathogenesis.

4.4. Anchoring to Lipids

A third mechanism of cell surface association is mediated by covalent linkage to lipids in the bacterial cell membrane (33). Prolipoproteins contain signal peptide sequences that differ from those in other secreted proteins, chiefly in the presence of a cysteine resi- due immediately following the site of cleavage. The thiol group in this cysteine is linked to an N-acyl diglyceride group of a glycerophospholipid by the action of an enzyme called prolipoprotein diacylglyceryl transferase (Lgt). After linkage of the cysteine, the signal pepide is removed by the type II signal peptidase, SPase II.

The EGDe genome encodes 68 predicted lipoproteins, representing approx 2.5% of all proteins, and approx 50% of all surface proteins (3,33). Twenty-four (~35%) of the lipoproteins are substrate-binding components of ABC transporter systems predicted to promote uptake of sugars, metals, amino acids, or peptides (3,52). OppA, a lipoprotein that mediates import of oligopeptides, plays a minor role in the growth of L. mono- cytogenes in host mammalian cells (53), suggesting that cytosolic bacteria might use peptides as a nutrient source. However, bacteria deleted for oppA are unaffected for viru- lence in mice, possibly because of the presence of multiple ABC transporters capable of importing di- or oligopeptides, or because host sugars can also be utilized as a food source (see Subheading 4.6.). The roles of other lipoprotein components of ABC trans- porters in pathogenesis have not been addressed. LpeA, a lipoprotein with homology to the PsaA adhesin of Streptococcus pneumoniae (54) was found to be required for internalization of L. monocytogenes into epithelial cell lines, but not for virulence in mice (55). In an effort to determine if lipoprotein maturation is essential for virulence, a gene (lsp) encoding a protein with homology to bacterial SPase II enzymes was iden- tified in the genomes of L. monocytogenes EGDe and L. innocua. Inactivation of lsp in L. monocytogenes resulted in a partial inhibition in maturation of LpeA, and a minor virulence defect in mice, perhaps because of an impairment in phagosomal escape (52).

Given the incomplete effect of the lsp mutation on SvpA processing, it is difficult to

conclude from these studies whether lipoprotein maturation is critical for L. monocyto-

genes pathogenesis. Importantly, a gene (lmo1101) encoding a second putative lipopro-

tein signal peptidase is present in the L. monocytogenes genome, but absent in L. innocua

(3). In future work, it would appear worthwhile to determine the effect of inactivation

of this gene, alone and in combination with lsp, on lipoprotein maturation and bacterial virulence. Another consideration is that removal of signal peptides might not be impor- tant for pathogenic properties of lipoproteins; perhaps it is only crucial that these pro- teins be anchored to the cytoplasmic membrane. Deletion of the L. monocytogenes lgt gene, encoding prolipoprotein diacylglyceryl transferase, might help determine the role of lipoprotein anchoring in virulence.

4.5. Electrostatic Anchoring of Proteins to Lipotechoic Acid

The entry-promoting protein, InlB, is bound to the bacterial surface through interac- tion of its positively charged carboxyl-terminal domain with negatively charged lipotech- oic acid present in the outer leaflet of the cell membrane (10,33). The InlB anchoring domain is comprised of three conserved, approx 80 amino acid “GW” modules con- taining the dipeptide glycine–tryptophan. Apart from InlB, EGDe contains seven other proteins with GW modules, with six of these also being present in L. innocua (3,33).

One of the species-conserved GW proteins, Ami, promotes adhesion to mammalian cells through its carboxyl-terminal GW modules (56). Ami contains an amino-terminal auto- lysin domain, which is also present in the related GW protein, Auto (57). Auto is absent from L. innocua, appears to aid InlA- and InlB-dependent entry into host cells, and is needed for efficient colonization of the intestine and liver of orally inoculated guinea pigs. Based on the findings with Ami and entry, it would be interesting to determine if the GW domain of Auto is sufficient for its role in mammalian cell invasion.

4.5.1. Export of Proteins to the External Environment (Secretion)

At least 105 proteins encoded in the EGDe genome are likely to be fully secreted, as judged by the presence of an SP sequence and the absence of a recognizable anchoring motif. This number represents about 4% of the L. monocytogenes proteome. The best characterized of the signal peptide-containing secreted proteins are the virulence fac- tors LLO, PlcA, PlcB, and a metalloprotease (Mpl) that mediates PlcB processing (1).

Another secreted protein demonstrated to have a role in virulence in mice is the LRR protein, InlC (31). The inlC gene is transcribed at low levels outside of host cells, and its expression is strongly induced on internalization of bacteria (31,58). inlC expres- sion is directly activated by the transcription factor PrfA (31,59).

Despite the intriguing data on InlC expression, the mechanism by which this protein affects pathogenesis is not clear. InlC is dispensable for phagosome escape, intracyto- solic replication, or cell-to-cell spread in all cell lines tested thus far (31,60). A recent report suggests that InlC might aid InlA-dependent entry into host cells (61). However, the fact that inlC expression is strongly upregulated inside host cells suggests that this LRR protein probably also influences postinternalization events. It is tempting to speculate that, like InlA and InlB, InlC may also promote virulence by interacting with one or more mammalian binding partners. In the case of InlC, the expectation is that at least one of its ligands is intracellular. The identification of InlC binding partners will likely provide important clues as to the mechanism(s) by which this protein affects pathogenesis.

In addition to InlC, LLO, PlcA, PlcB, and Mpl, the genome of EGDe encodes 18 other

secreted proteins that are absent from L. innocua (3) Presently, the role of these secreted

proteins in virulence is unknown. Interestingly, two of the uncharacterized secreted

proteins are members of the LRR family, raising the possibility that they could interact with host factors.

4.6. Proteins Involved in Nutrient Acquisition or Metabolism

A genomic approach was recently used to elucidate mechanisms by which L. mono- cytogenes obtains nutrients for replication in the mammalian cell cytosol (60,62). The intriguing observation that catabolism of hexose phosphates during growth in broth is PrfA-dependent (62,63) prompted a search for PrfA-regulated genes that might medi- ate sugar transport. A scan of the L. monocytogenes genome for open reading frames containing an upstream PrfA box led to the identification of a gene encoding a putative hexose phosphate transporter (Hpt) with sequence similarity to a known mammalian glucose-6-phosphate translocase (62). Hpt is absent from L. innocua. Subsequent expe- riments demonstrated that Hpt is needed for utilization of hexose phosphates (but not unphosphorylated sugars) in broth, efficient bacterial replication within host cells, and for virulence in mice. Taken together, these results provide compelling evidence that one or more host hexose phosphate(s) fuels cytosolic growth of L. monocytogenes.

Glucose-1-phosphate, the precursor and breakdown product of glycogen, is a plausible candidate for the growth-promoting sugar. Glycogen is abundant in the cytosol of hepa- tocytes (62,63), and hpt deletion mutants are defective in colonization of the liver of infected mice (62).

Fig. 2. Multiple levels of regulation of PrfA activity. At low temperatures (20–30qC) typically

encountered in the soil, PrfA translation is inhibited owing to occlusion of the ribosome bind-

ing site (RBS) in a stem-loop structure formed by the 5' untranslated region (UTR) of the mRNA

transcript originating from the P1 promoter or the upstream plcA promoter (not shown) (74). At

37qC, the secondary structure in the UTR is melted out, rendering the RBS accessible for transla-

tion. Transcripts directed from the P2 promoter lack the stem-loop structure; however, expression

from this promoter is low when PrfA is active (20). PrfA protein is regulated at the post-transla-

tional level by at least three different conditions. First, readily metabolizable, nonphosphorylated

sugars, including the plant-derived dissacharride, cellobiose, inhibit PrfA activity through an

undefined mechanism (20). Second, the presence of activated charcoal in the bacterial growth

medium enhances PrfA activity through sequestration of an unidentified bacterial “autorepres-

sor” (AR) (76). Finally, one or more host factors stimulate PrfA activity (78).

4.7. Transcriptional Regulators

The EGDe genome encodes 24 putative or known transcription factors that are absent from L. innocua CLIP11262 (3). Based on amino acid sequence, these transcriptional regulators can be placed into one of several families, including GntR, BglG, Xre, AraC, TetR, LysR, Fur, the CAP/Fnr family, and the response regulator (RR) proteins of two component-systems. Of all these proteins, only functions of the CAP/Fnr family mem- ber PrfA and the RRs CesR, LisR, and DegU have been determined. CesR, LisR, and DegU are all needed for efficient colonization of the spleen in intragastrically inoculated mice (64–66). CesR and its cognate histidine protein kinase (HPK), CesK, promote resistance to E-lactam antibiotics, and the virulence defect of the CesR/CesK mutants may reflect a role for this two-component system in controlling cell wall integrity dur- ing infection. LisR and its partner HPK, LisK, might also affect the cell wall, because these proteins influence resistance to cephalosporins and the lantabiotic, nisin (67). DegU is a member of a two-component system that promotes the development of genetic competence, degradative enzyme synthesis, and motility in the Gram-positive bacterium Bacillus subtilis (68). L. monocytogenes DegU is required for flagellin gene expression and motility (66). However, the fact that flagella are not required for virulence in mice (66) suggests that DegU promotes bacterial infections through a process distinct from motility. The transcription factor PrfA has been extensively studied since its discovery in 1991, and has been found to play a critical role in regulation of virulence gene expres- sion (20). Subheading 4.8. describes what is known about how the activity of PrfA is controlled, and Subheading 5. discusses results from a recent transcriptional profiling study that led to the identification of several classes of PrfA-controlled genes.

4.8. Regulation of PrfA Activity

The virulence genes hlyA (encoding LLO), plcA, mpl, actA, plcB, and inlC are all highly dependent on PrfA for their transcription (20,21). PrfA activity is upregulated on bacterial entry into host cells, resulting in a large induction in expression of these genes (20). Unlike the above genes, which have functions inside mammalian cells, the inlA and inlB genes are only partly dependent on PrfA for expression. inlA and inlB comprise an operon with multiple promoters, only one of which is controlled by PrfA (69,70). PrfA-independent expression of InlA and InlB makes intuitive sense, given that their role in internalization into host cells precedes PrfA activation in the cytosol.

PrfA has structural and functional similarity to the Catabolite Activator Protein (CAP)/

Fnr family of transcription factors (20,71,72). PrfA boxes located upstream of target pro-

moters conform to the consensus palindromic sequence TTAACANNTGTTAA, where

N is any nucleotide (20). The hly and plcA promoters have “perfect” consensus PrfA boxes,

whereas promoters for the remaining virulence genes have PrfA binding sites contain-

ing one or two base pair mismatches relative to the consensus. Promoters with perfect

consensus PrfA binding sites appear to have higher affinity for PrfA than those with

imperfect palindromes (71). In all PrfA-regulated target genes studied to date, the PrfA

box is centered approx 41 nucleotides upstream of the transcriptional start site, a situa-

tion that mirrors the position of CAP binding sites in so-called class II CAP-dependent

promoters (72). CAP-mediated transcriptional activation of class II promoters is known

to involve interactions between the transcription factor and the D and V

subunits of

RNA polymerase (RNAP) holoenzyme (72,73). It is possible that PrfA controls transcrip- tion through a similar mechanism.

PrfA is regulated by multiple mechanisms that allow increased transcriptional activ- ity in the infected host (Fig. 2). One mechanism is controlled by temperature (74). At low temperatures typical of growth in the soil (less than 30qC), translation of prfA mRNA is inhibited. This inhibition is probably because of occlusion of the ribosome binding site in a folded structure formed by the 5' untranslated region (UTR) of the prfA mRNA.

Higher temperatures encountered in host tissues (37qC) cause unfolding of the 5' UTR, resulting in access to the ribosome binding site and derepression of translation.

PrfA is also regulated at the post-translational level by several environmental, bacte- rial, and host-derived factors. Experiments with constitutively activated mutants of PrfA suggest that these factors control conformational changes that regulate the DNA binding activity of this transcriptional regulator (20,71,75). Readily metabolized, nonphosphor- ylated sugars including glucose, fructose, mannose, and E-glucosides, such as cellobiose, all strongly inhibit PrfA activity (20). In contrast, hexose phosphates, including the Hpt substrate glucose-1-phosphate, fail to impair PrfA activity (62,63). Interestingly, cellobiose is derived from plants and this disaccharide may serve the dual function of sustaining Listeria growth on decaying vegetation, while simultaneously preventing wasteful and inappropriate expression of PrfA-dependent virulence genes. Another mode of PrfA inhibition involves a diffusible molecule that is made by L. monocyto- genes and secreted into culture supernatants (76). This proteinaceous autorepressor (AR) can be removed by addition of activated charcoal to the growth medium. AR might regulate PrfA by direct interaction with the transcription factor or, alternatively, could antagonize the function of a putative PrfA-activating factor (77). Like cellobiose, AR could act to restrain PrfA activity during the saprophytic lifestyle of L. monocytogenes.

On internalization of individual bacteria into mammalian cell phagosomes and release into the cytosol, AR would likely be diluted, resulting in increased PrfA activity. Finally, the activation state of PrfA also appears to be controlled by one or more mammalian cell protein(s) (78). Host surface proteins could be involved, because upregulation of PrfA activity was found to require bacterial adhesion, but not internalization. Together, the multiple modes of regulation of PrfA ensure that the transcription factor is fully active only when L. monocyogenes infects host cells.

5. Use of Genomics or Proteomics to Characterize Changes in Bacterial Gene or Protein Expression

5.1. Identification of PrfA-Controlled Target Genes

Availability of the L. monocytogenes genome sequence has made it possible to ana- lyze transcription of all known or predicted open reading frames in response to specific genetic and/or environmental conditions. Buchrieser and colleagues used DNA arrays containing 99% of all predicted open reading frames in the EGDe genome to identify genes that are differentially expressed in wild-type and prfA-deletion mutant strains (21). Gene expression was monitored in the presence or absence of charcoal or cello- biose in order to assess the effect of conditions that enhance or inhibit PrfA activity.

Seventy PrfA-regulated genes organized in 47 predicted transcription units were iden-

tified. These genes fell into three distinct groups, based on whether PrfA augmented or

impaired transcription, and on the influence of charcoal or cellobiose on regulation.

Altogether, the results suggest the existence of at least two different activation states of PrfA, each of which regulates expression of distinct sets of target genes (Fig. 3).

Group I genes were directly upregulated by PrfA, in a manner enhanced by charcoal and diminished by cellobiose. This group contains key virulence genes already known to be under direct PrfA control, including hlyA, mpl, actA, plcA, plcB, inlA, inlB, inlC, hpt, and prfA itself. Two genes of unknown function, lmo2219 and lmo0788, also fall into group I. Based on the presence of upstream PrfA boxes and the distance between these binding sites and the predicted –10 promoter elements, it seems likely that lmo2219 and lmo0788 are also directly regulated by PrfA.

PrfA negatively regulates expression of group II genes (Fig. 3). Interestingly, PrfA- mediated repression was insensitive to charcoal or cellobiose, suggesting that regula- tion of these genes may not require fully active PrfA. Eight group II genes comprising two putative transcription units were identified. One transcription unit contains seven genes (lmo0178–0184), some of which encode substrate-binding and permease com- Fig. 3. Three classes of PrfA-regulated genes. Three categories of PrfA-controlled genes were identified in DNA microarray study performed by Buchrieser and colleagues (21). Group I genes include nine well-characterized virulence genes whose expression was previously known to be promoted by PrfA, as well as two novel genes (lmo2219 and lmo0788) of undefined func- tion. Expression of these genes is enhanced by charcoal and impaired by cellobiose, suggesting that a high activity form of PrfA (PrfA*) directs their transcription. Surprisingly, cellobiose did not alter expression of the two group II and 35 group III transcription units identified in the study.

These results raise the possibility that a low activity form of PrfA might direct transcription of

these genes. Expression of group II genes was repressed by PrfA, whereas transcription of group

III genes was enhanced. “Direct?” indicates the presence of an upstream PrfA box, and the possi-

bility the PrfA directly regulates gene expression. “Indirect” indicates that absence of a predicted

PrfA-binding site making it improbable that the transcription unit is directly controlled by PrfA.

ponents of an ABC transport system predicted to import disaccharides, oligosaccharides, and polyols. The second transcription unit, lmo0278, is monocistronic and codes for the ATPase component of a sugar transporter. It is possible that the products of the two group II transcription units act together to mediate sugar import. lmo0278 has a poten- tial PrfA box located only about four nucleotides upstream of its –10 region. The prox- imity of these binding sites to the –10 might inhibit transcription by preventing binding of RNAP. In contrast to the situation with lmo0278, lmo0178 lacks an upstream PrfA box near predicted –10 or –35 regions of the promoter, and the mechanism by which PrfA downregulates expression of this gene is unclear. It is also not obvious why it would be advantageous for L. monocytogenes to impair expression of this transport system through PrfA. One possible explanation might be that the group II genes encode a transporter of sugars that are readily available in the host cell cytosol and capable of inhibiting PrfA activity. In this case, repression of these genes would be crucial for the ability of PrfA to efficiently activate virulence gene expression in mammalian cells. This idea could be tested by identification of the specific carbohydrates transported by the group II gene products and determining the effects of these sugars on PrfA activity.

Group III genes were positively regulated by PrfA, in a manner that was insensitive to cellobiose (Fig. 3). Again, regulation of these genes might not require full PrfA activity.

Fifty-three class III genes organized in 37 transcription units were found. Two of these genes, bsh and lmo0596, contain potential PrfA binding sites located 18–24 nucleo- tides upstream of the –10 regions of their promoters (3,79). Thus, these genes are likely to be direct targets of PrfA. bsh codes for a bile salt hydrolase that plays an important role in virulence, presumably by protecting L. monocytogenes from toxic effects of bile salts in the duodenum and liver (79). The recent and unexpected finding that L. mono- cytogenes colonizes the gall bladder in mice (80) suggests that bsh also might promote bacterial survival in this bile-producing organ. The function of lmo0596 is unknown.

In contrast to the situation with bsh and lmo0596, the vast majority of group III genes

lack upstream PrfA boxes. It seems likely that PrfA induces expression of one or more

intermediary factors that then exert more direct effects on transcription (see Subhead-

ing 5.2.). The functions of about 70% of the indirectly regulated group III genes have

been characterized or predicted based on sequence (Fig. 3). One group of proteins that

have been extensively studied is the OpuC carnitine/betaine ABC transport system,

which promotes growth in high salt conditions (osmotolerance) and is also needed for

colonization of the small intestine and livers of infected mice (81). OpuC might pro-

mote growth and survival of L. monocytogenes in the high salt/low water environment

of many food products and also in the duodenum, which has an osmolarity equivalent to

approx 0.3 M NaCl (81). The predicted functions of the remainder of the group III pro-

teins include regulation of the cellular redox environment (oxidoreductase- Lmo0669),

di- and/or tripeptide transport (Lmo0555), pyruvate-dependent transport system-medi-

ated import of mannose, uptake of glucose or other sugars (Lmo169), and carbon cata-

bolism (succinate semi-aldehyde dehydrogenase- Lmo0913; dihydroxyacetone kinase

subunits- Lmo2695-2697). Although not identified in the microarray study, another tran-

scription unit that may be classified as belonging to group III is the bilE operon (82). This

operon appears to be indirectly regulated by PrfA, and encodes a two-component bile

exporter that promotes resistance to bile in vitro and intestinal colonization in mice. Taken

together, the findings suggest that PrfA-mediated expression of group III genes may help

L. monocytogenes cope with toxicity associated with osmotic shock, oxidative stress, and bile salts, and perhaps also provide energy for growth in the mammalian cytosol.

5.2. Interaction Between the PrfA and V V V V V

Regulons

The DNA array study of PrfA-regulated genes revealed an unexpected interaction between regulons controlled by PrfA and the alternative sigma factor, V

. In L. mono- cytogenes, Bacillus subtilis, and other Gram-positive bacteria,V

is activated and pro- motes survival during multiple stress conditions, including starvation, acid, osmotic, heat, or oxidative shock (83–86). The group I inlAB operon, all of the group II transcription units, and about 60% (22/37) of the group III transcription units, including the opuC and bilE operons, are predicted to contain promoters transcribed by RNAP containing V

(21). Moreover, a DNA microarray study comparing gene expression in wild-type and sigB deletion strains of L. monocytogenes demonstrated that inlAB, bsh, opuC, bilE, and at least five of the other group III transciption units indicated in Fig. 3, are indeed controlled by V

(82,87). Experiments with a sigB null mutant strain support the idea that V

-dependent expression of inlA, bsh, and opuC is critical for the function of these genes (79,83,85,88–91). Finally, one of the three promoters (P2) known to drive PrfA expression has consensus V

–10 and –35 sites (Fig. 4A), and is transcribed by V

-con- taining RNAP in vitro (92). V

-mediated transcription of PrfA from P2 could, at least in part, be responsible for the stimulatory effects of oxidative stress, high osmolarity, and severe heat shock (42–48qC) on PrfA activity (20).

For genes with upstream PrfA boxes, PrfA and V

are likely to regulate transcription by acting at different promoters. Both inlAB and bsh contain multiple promoters (69,79), and sequence information suggests that at least one promoter for each of these tran- scription units is probably transcribed by RNAP containing the major sigma factor, V

(Fig. 4B). Based on spacing between –10 sites and PrfA boxes, it seems likely that PrfA activates expression of the V

promoters for both inlAB and bsh, and that the V

pro- moters are not direct PrfA targets (Fig. 4B). The fact that PrfA has no influence on V

- directed transcription of bsh in vitro is consistent with this idea (92). The existence of separate promoters for V

and PrfA could provide a mechanism for distinct environ- mental signals to regulate gene expression at different steps of infection. In the mildly acidic conditions (pH 4.5–6.5) and elevated osmolarity (~0.3 M NaCl) of the small intestine, V

-dependent transcription of bsh and/or inlAB could play an important role in acquisition of resistance to bile salts and/or enhancement of bacterial invasion. V

is probably uniquely suited to boost gene expression at this early step in infection, as PrfA is inhibited by low pH (20). The contribution of PrfA to inlAB and bsh transcrip- tion might become more important during infection of the liver, where the pH of fluids is closer to neutral. It would be interesting to determine the relative contributions of the different promoters upstream of inlAB and bsh to intestinal and hepatic coloniza- tion using orally inoculated guinea pigs and/or transgenic mice expressing human E- cadherin (23).

A central unresolved question is how PrfA promotes expression of the majority of

group III genes, which lack identifiable upstream PrfA-binding sites. Three different

models are presented in Fig. 4C. In two of the models, PrfA activates expression of a gene

whose product (X) directly stimulates transcription at either V

- and/or V

-dependent

promoters. In the third model, factor X stimulates V

activity. A particularly attractive

candidate for “X” in model 3 is RsbV, an anti-antisigma factor that positively regulates V

activity in a variety of Gram-positive bacteria (86). rsbV was identified as a group III gene in EGDe but, surprisingly, PrfA did not enhance rsbV expression in a different L. monocytogenes strain that has an almost identical profile of group III genes (21).

Hence, rsbV transcription alone cannot account for the PrfA-dependency of group III

genes. Moreover, rsbV lacks upstream PrfA boxes, indicating that it is unlikely to be

Fig. 4. Interaction between the PrfA and VB regulons (A) PrfA transcription. The three pro-

moters directing PrfA expression are depicted. The # symbol indicates that PrfA directly acti-

vates transcription of the plcA promoter by binding to an upstream PrfA box. Activated PrfA

(PrfA*) and RNA polymerase (RNAP) holoenzyme containing V

(EVB) control expression of

distinct promoters. EVA indicates that transcription of the plcA and P1 promoters is directed by

RNAP containing the major sigma factor, V

. It is worth noting that data in a recent report sug-

gests that P2 is actually comprised of two overlapping promoters, one of which is transcribed

by V

-containing RNAP and the other of which is recognized by V

(92). For the sake of sim-

plicity, only V

-dependent transcription is depicted. (B) Transcription of inlAB and bsh. The

roles of PrfA and V

in controlling gene expression is illustrated. Similar to the situation with

the plcA–prfA operon, PrfA and V

appear to regulate gene expression by acting at distinct pro-

moters. The fact that bsh expression is insensitive to cellobiose (21) suggests that PrfA might

not need to be fully active in order promote transcription of the gene. (C) Group III genes

indirectly regulated by PrfA. PrfA is predicted to stimulate expression of one or more genes (x)

that have more direct effects on group III gene expression. x could represent a single gene that

controls transcription f all group III genes or, alternatively, several x genes could exist, each of

which regulates expression of distinct subsets of group III transcription units. Three potential

modes by which X might indirectly control gene expression are presented. X could be a tran-

scription factor that activates transcription from V

- (1) or V

- (2) dependent promoters. Alter-

natively, X might enhance V

activity (3).

directly controlled by PrfA. Clearly, understanding the mechanism(s) by which PrfA affects group III gene expression will require identification of factor X. This could be accomplished by genetic screens to identify mutations that simultaneously affect expres- sion of several group III genes. In addition, biochemical studies might lead to the identi- fication of factor(s) that bind group III promoters, providing evidence for models 1 or 2.

6. Future Challenges of Genome-Based Approaches

Although the complete sequence of the L. monocytogenes genome has been available for less than 4 yr, it is already clear that genomics has made a large impact on Listeria research. Several studies have utilized genome sequence information as an indispensa- ble tool to provide clues about whether a particular gene might be involved in virulence.

Approaches include mutating L. monocytogenes genes that are absent in L. innocua and/or have sequence similarity to genes of known function from other bacteria, and the use of DNA microarrays to identify bacterial genes that might affect Listeria patho- genesis. In addition, although not discussed in this review, several recent proteomic studies have led to the identification of bacterial proteins that are secreted or differen- tially expressed under particular environmental conditions, such as starvation or bio- film formation (41,90,93–96). On the side of the host organism, information on the human and mouse genomes has allowed the study of mammalian genes and proteins that may affect susceptibility to Listeria (97–100).

The work outlined in this chapter represents the first generation of studies utilizing the power of genomics. Future work is likely to expand and improve on the initial studies in several respects. For example, transcriptional profiling experiments that more closely mirror in vivo physiological situations could be performed by analysis of mRNA iso- lated from infected host cells or mice. The use of bacterial strains deleted for genes encoding PrfA or V

could confirm, extend, or modify regulons that had been identi- fied using artificial or simplified in vitro activating or repressing conditions, like growth in charcoal or cellobiose. Simultaneous analysis of host and bacterial gene expression through DNA microarrays has recently been performed in mice infected with a patho- genic E. coli strain (101), and similar experiments could probably be performed with L. monocytogenes. Such studies could lead to identification of bacterial or host genes that are differentially expressed in various tissues, such as the intestinal epithelium, the liver, or the spleen.

Another likely future direction in Listeria genomics is the analysis of bacterial genes

that have been previously neglected because the sequence of their protein products did

not suggest a precise molecular function. One category of “neglected” genes is those

whose proteins have signature motifs that allow them to be confidently placed into

general functional classes, such as transcription factors, two-component HPKs or RRs,

or LRR proteins. Although sequence information predicts that the corresponding pro-

teins are likely to regulate transcription or promote protein–protein interactions, their

exact responses and molecular targets are difficult if not impossible to divine. It will be

important to systematically and exhaustively examine the roles of these proteins using

functional genomic approaches that are currently being utilized to great effect with the

budding yeast Saccharomyces cerevisiae (102). For example, the functions of unchar-

acterized HPKs, RRs, or other transcription factors can begin to be probed by transcrip-

tional profiling using DNA microarrays of the L. monocytogenes genome. In the case

of RRs and other DNA binding proteins, the approach can be modified so as to identify target genes that are under direct control of the transcriptional regulator by performing chromatin immunoprecipitation (ChIP) of the DNA binding protein of interest, followed by hybridization to DNA microarrays (chip). This “ChIP-on-chip” approach has been recently employed to identify direct targets of DNA binding proteins in yeast (102) and bacteria (103,104). Finally, the near universal involvement of LRR domains in pro- tein–protein interactions (34) suggests that the function of an uncharacterized Listeria LRR protein can be effectively investigated through comprehensive analysis of protein complexes. This goal could be accomplished through two-hybrid-based screens of geno- mic libraries using the LRR protein as bait, or affinity purification of a tagged LRR protein followed by identification of copurifying ligands through mass spectrometry.

Two-hybrid or mass spectometry strategies have been successfully used to comprehen- sively characterize protein complexes in the Gram-negative bacterial pathogen Helico- bacter pylori (105) and in yeast (102). However, it is vital to consider that the two L.

monocytogenes LRR proteins that have been most extensively studied, InlA and InlB, bind to mammalian ligands. The eleven EGDe-specific LRR proteins whose binding partners have yet to be identified (3) could have bacterial and/or host binding partners.

Hence a thorough investigation of a given LRR protein will require attempts to isolate binding partners from both the bacterial and mammalian proteomes.

Another future goal will be to use the genomes of serotype 4b epidemic strains to identify bacterial factors that may enhance pathogenic potential. Such factors might be present in 4b strains, but absent in other L. monocytogenes serotypes. One potential complication is that existing mouse or guinea pig animal models may not be appropriate for identification of auxillary virulence factors that augment pathogenicity in humans.

Recent advances in inhibition of mammalian gene expression through RNA interfer- ence (RNAi) technology suggest that it may become possible in the not-too-distant future to perform genome-wide screens to identify mouse or human genes that regulate various aspects of Listeria pathogenesis. A recent report describes a phenotypic screen using microarrays containing spots of mammalian cells in which expression of various target genes were “knocked down” through transfection with small interfering RNAs (siRNAs) (106). Although only seven genes were targeted in the study, plans to assem- ble comprehensive RNAi collections for the human genome will likely make large- scale genetic screens a reality (107). siRNA–mammalian cell microarrays could be used to elucidate many important events in Listeria pathogenesis. For example, a recent study described an innate immune response induced by cytosolic L. monocytogenes that results in induction of the E interferon gene and downstream interferon responsive genes (100).

When comprehensive RNAi microarrays are available, these could potentially be em-

ployed with a mammalian cell line expressing a luciferase reporter driven by the E inter-

feron promoter to identify host genes that sense or transduce signals elicited by cytosolic

bacteria. Similarly, siRNA-mammalian cell arrays could be infected with L. monocyto-

genes harboring a transcriptional reporter of a PrfA-controlled target gene, with the goal

of finding host genes that stimulate PrfA activity (78). In addition to cell microarrays

containing siRNA, microarrays in which a cell line is subjected to a battery of chemical

inhibitors have also been described (106,107). In principle, such arrays could be used to

screen natural or synthetic compounds for the ability to impair key events in Listeria

virulence, such as bacterial internalization or actin-dependent movement.

Although several mammalian genes that affect susceptibility to L. monocytogenes infection have been identified through gene targeting (8), transgenics (23), or other approaches (108), the list of host susceptibility factors is far from complete. Recently, genome-wide chemical mutagenesis of mice has been performed with the aim of per- forming unbiased screens for a variety of novel mutant phenotypes (109,110). Efforts are underway to screen mouse mutants for susceptibility to L. monocytogenes and other bac- terial pathogens ([111] http://www.mikrobio.med.tu-muenchen.de/forschung/enu.html).

Finally, in addition to animals or mammalian cell lines, other model systems might prove useful in genomic-based investigations of host susceptibility factors. For exam- ple, L. monocytogenes establishes infections in Drosophila melanogaster cell lines or whole organisms, in a manner that depends on LLO-mediated escape and ActA-depen- dent cell–cell spread (112,113). The availability of Drosophila mutants and the amena- bility of Drosophila cell lines and animals to RNAi-mediated gene silencing may make this metazoan useful for identification of host proteins that play an evolutionarily con- served role in defense against intracellular pathogens. Clearly, genomic-based strategies will play an increasingly important role in investigations of mechanisms of virulence of L. monocytogenes and other bacterial pathogens. Of course, equally important is the use of classic genetic and biochemical approaches to thoroughly probe the function of any candidate virulence factor identified from genomics.

References

1. Vazquez-Boland, J. A., Kuhn, M., Berche, P., et al. (2001) Listeria pathogenesis and molec- ular virulence determinants. Clin. Microbiol. Rev. 14, 584–640.

2. Mead, P. S., Slutsker, L., Dietz, V., et al. (1999) Food-related illness and death in the United States. Emerg. Infec. Dis. 5, 607–625.

3. Glaser, P., Frangeul, L., Buchrieser, C., et al. (2001) Comparative genomics of Listeria species. Science 294, 849–852.

4. Tettelin, H. E., Saunders, N. J., Heidelberg, J., et al. (2000) Complete genome sequence of Neisseria meningitidis serogroup B strain MC58. Science 287, 1809–1815.

5. Parkhill, J., Achtman, M., James, D. D., et al. (2000) Complete DNA sequence of a sero- group A strain of Neisseria meningitidis Z2491. Nature 404, 502–506.

6. Subtil, A. and Dautry-Varsat, A. (2004) Chlamydia: five years A.G. (after genome). Curr.

Opin. Microbiol. 7, 85–92.

7. Dussurget, O., Pizarro-Cerda, J., and Cossart, P. (2004) Molecular determinants of Listeria monocytogenes virulence. Annu. Rev. Microbiol. 58, 587–610.

8. Pamer, E. (2004) Immune responses to Listeria monocytogenes. Nat. Rev. Immunol. 4, 812–823.

9. Tilney, L. G. and Portnoy, D. A. (1989) Actin filaments and the growth, movement, and spread of the intracellular bacterial parasite Listeria monocytogenes. J. Cell Biol. 109, 1597–1608.

10. Cossart, P., Pizarro-Cerda, J., and Lecuit, M. (2003) Invasion of mammalian cells by Lis- teria monocytogenes: functional mimicry to subvert cellular functions. Trends Cell Biol.

13, 23–31.

11. Schubert, W. D. and Heinz, D. W. (2003) Structural aspects of adhesion to and invasion of host cells by the human pathogen Listeria monocytogenes. Chembiochem. 4, 1285–1291.

12. Mengaud, J., Ohayon, H., Gounon, P., Mège, R. M., and Cossart, P. (1996) E-cadherin is

the receptor for internalin, a surface protein required for entry of Listeria monocytogenes

into epithelial cells. Cell 84, 923–932.

13. Shen, Y., Naujokas, M., Park, M., and Ireton, K. (2000) InlB-dependent internalization of Listeria is mediated by the Met receptor tyrosine kinase. Cell 103, 501–510.

14. Cossart, P. and Sansonetti, P. J. (2004) Bacterial invasion: the paradigms of enteroinva- sive pathogens. Science 304, 242–248.

15. Sun, H., Shen, Y., Dokainish, H., Holgado-Madruga, M., Wong, A., and Ireton, K. (2005) Host adaptor proteins Gab1 and CrkII promote InlB-dependent entry of Listeria mono- cytogenes. Cell Microbiol. 7, 443–457.

16. Portnoy, D. A., Auerbuch, V., and Glomski, I. J. (2002) The cell biology of Listeria mono- cytogenes infection: the intersection of bacterial pathogenesis and cell-mediated immu- nity. J. Cell Biol. 158, 409–414.

17. Marquis, H., Doshi, V., and Portnoy, D. A. (1995) The broad-range phospholipase C and a metalloprotease mediate listeriolysin O-independent escape of Listeria monocytogenes from a primary vacuole in human epithelial cells. Infect. Immun. 63, 4531–4534.

18. Smith, G. A., Marquis, H., Jones, S., Johnston, N. C., Portnoy, D. A., and Goldfine, H.

(1995) The two distinct phospholipases C of Listeria monocytogenes have ovelapping roles in escape from a vacuole and cell-to-cell spread. Infec. Immun. 63, 4231–4237.

19. Gedde, M. M., Higgins, D. E., Tilney, L. G., and Portnoy, D. A. (2000) Role of Listerio- lysin O in cell-to-cell spread of Listeria monocytogenes. Infect. Immun. 68, 999–1003.

20. Kreft, J. and Vazquez-Boland, J. A. (2001) Regulation of virulence genes in Listeria. Int.

J. Med. Microbiol. 291, 145–157.

21. Milohanic, E., Glaser, P., Coppee, J.-Y., et al. (2003) Transcriptome analysis of Listeria monocytogenes identifies three groups of genes differently regulated by PrfA. Mol. Micro- biol. 47, 1613–1625.

22. Drevets, D. A., Leenen, P. J. M., and Greenfield, R. A. (2004) Invasion of the central nervous system by intracellular bacteria. Clin. Microbiol. Rev. 17, 323–347.

23. Lecuit, M., Vandormael-Pournin, S., Lefort, J., et al. (2001) A transgenic model for Lister- iosis: role of internalin in crossing the intestinal barrier. Science 292, 1722–1725.

24. Dramsi, S., Biswas, I., Maguin, E., Braun, L., Mastroeni, P., and Cossart, P. (1995) Entry of L. monocytogenes into hepatocytes requires expression of InlB, a surface protein of the internalin multigene family. Mol. Microbiol. 16, 251–261.

25. Gaillard, J. L., Jaubert, F., and Berche, P. (1996) The inlAB locus mediates the entry of Listeria monocytogenes into hepatocytes in vivo. J. Exp. Med. 183, 359–369.

26. Join-Lambert, O. F., Ezine, S., Le Monnier, A., et al. (2005) Listeria monocytogenes infected bone marrow myeloid cells promote bacterial invasion of the central nervous system. Cell Microbiol. 7, 167–180.

27. Lecuit, M., Nelson, D. M., Smith, S. D., et al. (2004) Targeting and crossing of the human maternofetal barrier by Listeria monocytogenes: Role of internalin interaction with tro- phoblast E-cadherin. Proc. Natl. Acad. Sci. USA 101, 6142–6157.

28. Bakardjiev, A. I., Stacy, B. A., Fisher, S. J., and Portnoy, D. A. (2004) Listeriosis in the pregnant guinea pig: a model of vertical transmission. Infect. Immun. 72, 489–497.

29. Dramsi, S., Dehoux, P., Lebrun, M., Goossens, P. L., and Cossart, P. (1997) Identification of four new members of the internalin multigene family of Listeria monocytogenes EGD.

Infect. Immun. 65, 1615–1625.

30. Raffelsbauer, D., Bubert, A., Engelbrecht, F., et al. (1998) The gene cluster inlC2DE of Listeria monocytogenes contains additional new internalin genes and is important for virul- ence in mice. Mol. Gen. Genet. 260, 144–158.

31. Engelbrecht, F., Chun, S.-K., Ochs, C., et al. (1996) A new PrfA-regulated gene of Listeria

monocytogenes encoding a small, secreted protein which belongs to the family of inter-

nalins. Mol. Microbiol. 21, 823–837.